DØ Level 3 Algorithms Meeting:
6th February 2002
Talks
-
Matching MC/L1/L2/L3/RECO Physics Objects:
Reiner Hauser
The idea is to provide a common format for describing physics objects
at each level and to provide matching between those objects.
-
Certification of p11.01.00 JET and
Electron ID: Volker Buescher
The conclusion was that JET and
Electron ID are in excellent shape in p11 and are ready to be used
online.
Calculation of the EM fraction within a narrower (delta_r=0.2) cone
looks like it improves the performance, but more tests needed before
making this the default.
-
Shower Shape for Electron ID: Ulla Blumenschein
Shower shape variables based on an energy-weighted delta_r of the
cluster increase the rejection factor to around the factor of 10
needed for p11.
Some code will need to go in as a "bug fix" to allow this to be used
in a future version of p11.
-
Tau ID: Yann Coadou
Plots for p11.01.00 out of the box were shown that look very similar
to the already running p10 code.
100K events were run (using JET, EM and tau tools) without a crash.
A bug has been found in the tau filter script: numbers of objects were
required to be greater than rather than
the desired greater than or equal to
the cut value specified in the trigger list.
The bug arose because the filter script was not tested directly, but
instead the filter was simulated by applying cuts at the roottuple
level.
For future releases we should make sure that this stage is properly
included in the testing.
Conclusion on p11 Certification
It would be highly desirable for p11 to be remade with
the tau filter bug fixed (p11.01.01?) and released to the farm according to the
originally forseen 11th February schedule.
The question arose (and was not entirely resolved) as to whether or
not the relevant ID groups have to "rubber stamp" this p11 certification
before it can be run online.
Verbal Status Report
-
L3Muon Status: Martijn Mulders and Martin Wegner.
The main problem encountered in the last week concernes getting the
correct unpacking of the muon information in data and Monte Carlo.
It was concluded that:
-
The real solution to these problems is to use the muon unpacker
written by Scott Snyder, which figures out the configuration of the
system from the data.
Dennis Shpakov committed himself to certifying this package before
Christmas and this work is urgently needed.
-
Obtaining correct cfg.dat files needed for the old unpacker has proven
to be more difficult than expected.
In the short term this needs to be solved so that
certification of p11 can be re-started.
It has been shown that a
cfg.dat for a particular data run can be successfully generated
automatically using a utility program, but this doesn't appear to work
for Monte Carlo. Why doesn't the default cfg.dat file work for Monte Carlo?
Points Arising, AOB.
-
Jon will consult with Gustaaf and Alan Jonckheere wrt making p11 update.
-
Scribe: Terry Wyatt.